Scholarlink

<p>Recognizing signs from the Arabic alphabet facilitates communication and enhances communication accessibility for individuals utilizing sign language due to deafness or hearing impairment. Allowing them to express themselves and understand more effectively, thereby promoting more inclusive communication. This paper introduces an approach that effectively tackles the problem of Arabic sign language recognition by combining the capabilities of two models: EfficientNet B0 and ConvNeXtSmall. Extensive comparative analyses reveal that the new combined model achieves better performance than the competitive models in terms of accuracy and overall performance in the classification of Arabic alphabet signs.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Efficient convolutional neural architecture for Arabic Sign Alphabet Classification

  • 期刊论文

  • 发表日期：

    2025年11月10日

查看全文
  • Sarra Rouabhi,
  • Abdennour Azerine,
  • Redouane Tlemsani,
  • Lhassane Idoumghar

摘要

Recognizing signs from the Arabic alphabet facilitates communication and enhances communication accessibility for individuals utilizing sign language due to deafness or hearing impairment. Allowing them to express themselves and understand more effectively, thereby promoting more inclusive communication. This paper introduces an approach that effectively tackles the problem of Arabic sign language recognition by combining the capabilities of two models: EfficientNet B0 and ConvNeXtSmall. Extensive comparative analyses reveal that the new combined model achieves better performance than the competitive models in terms of accuracy and overall performance in the classification of Arabic alphabet signs.